Automated reasoning systems such as theorem provers often employ interaction or cooperation with further reasoning software. Whereas in most cases the concrete choice of cooperating software is, to some extent, irrelevant, these systems are nevertheless often rigid in practice due to compatibility issues. In order to support more flexible cooperation schemes, a machine-readable description format for automated reasoning systems' capabilities is proposed. Additionally, a simple HTTP-based protocol for system and capability discovery is outlined. Both the format and the protocol are designed to be simple, extensible and easy to use with none to minor modifications for existing reasoning systems.